CAM, Assembly and Simulation

[topcyde]

I was curious if the assembly workbench and the CAM workbench will be compatible? My concern is that I'd like to eventually be able to build a machine in the Assembly work bench,and use it to drive the simulation and toolpath verification in the CAM workbench. I presently do this with my integrated commercial CAD/CAM system and was trying to figureout how these are tied together in FreeCAD. Maybe my process is too well defined by present experience, but I'm not seeing how this would work. Could someone perhaps explain if it would work and how?

[aj5robinson]

Hi Topcyde,

At this point, not much thought has gone into integration between the Cam module and the others apart from the following discussion: viewtopic.php?f=15&t=5802

Our current focus is getting basic functionality up and running but we seem to be very time-poor of late. I (we) are happy to hear your experiences with other tools and ideas for how it could work (hopefully better).

Currently the workflow for the cam module is (or planned to) be:
0) create model (in other workbenches)
1) generate toolpath using 1 or more TPGs (this is python syntax like HeeksCNC)
2) post process this toolpath using a post-processor suitable for your target machine (this is GCode or other syntax)
3) parse machine code back into internal format for rendering as lines over model (lines can be selectively displayed).

From here it could be possible to implement a simple simulator that moves a tool model along this tool path (like pycam) or something more advanced such as using a robot or other feature within another workbench; all this hasn't really been planned yet so plenty of opportunity to get your ideas in there.


Cheers
Andrew

[topcyde]

I had read the discussion you linked to. I really think the direction from that post is pretty much on point. The only thing I didn't see mentioned in it's role there was postprocessing, this has always been the hiccup in CAM. Getting machine usable code out of the computer with out hours of hand edit afterwards. This is where your Step 3 from above comes most into play. Good simulation means better code. Toolpath generators are important but once people figure out that they can solve trochoidal toolpaths using a little imagination and a vector overlay of cut region, Toolpath generation becomes moot. The bestest fastes toolpath is utterly useless if the machine can't run it and you waste a colossal amount of time, tooling, and machine repair costs, without good simulation(although we call that experience!).

When this gets far enough I'd like to help build and test simulations. I have access to machines with multiple controls, and multiple configurations. From stupid 18axis multi spindle mill turns, 5axis mills, turret lathes, and just about one of everything in between, but I've also got a linuxcnc control at home and a couple of friends with machines in the works.

Generating machine code is easy. WIth a little experience most machine programmers should be able to hand write simlutaneous 4axis code. Simulation sure is handy though for those times you don't realize you're thinking on the wrong side of the coordinate system.